For adding a new road to a map data used in a navigation apparatus, a map manufacturer edits and adds a road data with map edit software and performs a full update or an increment update of the map data. For this updating, measurement vehicles dedicated to the map manufacturer travel and collect a road data, and thereafter, the map manufacturer makes a map data for several months and then distributes the map data to users. Thus, it takes a long time to collect the road data and create the map data. This long time degrades freshness of the ready-made map data because of a real-world change in road feature.
In this relation, a proposed navigation apparatus learns a new road unregistered in the ready-made map data based on a movement trace of a movable body such as a vehicle and the like (see Patent Documents 1 and 2). A road learning function generates and records a new road (a learned road) absent in the ready-made map data, based on a movement trace from a point where present position of the movable body departs from a prestored road in the map data to a point where the present position of the movable body returns to the prestored road in the map data.
Patent Document 1: JP-H6-88733A
Patent Document 2: JP-2006-125883A
In the road learning, it may be preferable to add a learned road data to a ready-made map data without changing the ready-made map data. This is because once the ready-made map data is changed, it becomes difficult to ensure map data integrity when the map manufacturer performs the incremental update of the ready-made map data.
However, when both of the prestored road in the ready-made map data and the learned road are present, a guidance route different than an actual road condition may be set depending on a data structure associating the prestored road and the learned road.